LUND INSTITUTE OF TECHNOLOGY

LUND UNIVERSITY
Division of Building Materials

lsothermal calorimetry for the

study of cement hydration

Lars Wads

wBM-7166

Lund 2001

Isothermal calorimetry for the study of cement

hydration
Lars \tVads

August 28,2007

Summary

Isothermal heat conduction calorimetry is a useful method for determining the

heat production rate and the developed heat during cement hydration. It may be
used for several different purposes: 1. To determine the total heat ofhydration.
2. To determine the rate of heat production in concretes. 3. As a general tool
to investigate the hydration process. Flom the studies presented in this report
it is concluded that the heat production rate per mass of cement is quite similar
for cement pastes and cement mortars (and then probably also for concretes) of
the same water-cement ratio. Also, the way that a cement paste is mixed does
not much influence the reaction rate.
Two measurement procedures that are alternatives to solution calorimetry
and (semi-)adiabatic calorimetry are presented in this report. The methods
described are similar to the Cembureau Test Method of 1977. Both methods
use rather small hand mixes and the small samples are quickly charged into the
calorimeter to be able to measure the thermal power as early as possible from
the start of the reaction.

2 fntroduction
This is the result report from the fi.rst part of NORDTEST project 1534-01 to
develop a method based on isothermal heat conduction calorimetry to measure
the heat of hydration of cement, mortar and micro concrete. This method is to
be ar alternative to the calorimetric methods in use.
In the nordic cement industry there are principally two calorimetric methods
used to determine heat of hydation of Portland cement: the solution calorimeter
and the (semi-)adiabatic calorimeter. For the solution calorimeter there exists
a number of standards (DIN 1164-8, ASTM C186-98, prEN 196-8, SS 81 1960

etc.) and for the semi-adiabatic calorimeter there are Flench standards (NFP
15-436) and Brittish standards (BS 4550). There has also been NORDTEST
projects using semi-adiabatic calorimetry (NT Build 388 and NT Build 480).
1

This report has a rather wide scope. It both gives a general background
of the usefulness of isothermal heat conduction calorimetry for the study of
cement hydration, describes procedures for measurements on cement hvdration,
and discusses calibration and other important issues in a number of appendices.
The measurement methods proposed will be used in the second part of this
project to compare isothermal calorimetry to solution calorimetry and semiadiabatic calorimetry. 1
A note for quick-readers:

Sections 4 and 5 give the proposed measurement methods for isothermal

calorimetry to be used in the second part of this project.

o The first appendix gives a list of possible uses of isothermal calorimetry

in cement science.
o The rest of the appendices

discusses various aspects of isothermal calorime-

try-

Principle of isotherrnal heat conduction-calorime-

try
The method comprises the measurement of thermal power (heat production
rate) and specifi.c thermal power (heat production rate per gram cement) of the
cement hydration reaction at essentially constant temperature. The primary
result of a mesurement with an isothermal heat conduction calorimeter is the
thermal power as a function of time. This result is often rather detailed and one
can usually pinpoint the exact time when the main reaction starts, find multiple
peaks etc. Because of this, isothermal calorimetry can have many different uses,
some of which are described above.
The instrument used is an isothermal heat conduction calorimeter. In such
an instrument the heat produced in the sample is conducted away from the sample so that the sample stays at essentially isothermal conditions. The thermal
power is measured by a heat flow sensor.
The heat production rate is not as dependent on water-cement ratio, mixing
technique and aggregate as are ma,ny other measurement techniques used for
cement, mortar and concrete. These issues can therefore be treated in less
detail than done in ES 196-1 and other standards. This is further discussed in

Appendix 3 and

4.

lReferences to the use of the TAM Air (Thermometric AB, Jirfll,

Sweden) isothermal
heat conduction calorimeter will be made in footnotes. It is a state-of-the at instrument
suitable for cement hydration studies that has been used in both the fist and the second
parts of this study. Note that the TAM Air is not suounded by any patents and that the
general technique is well known and has been used in cement science for nearly a centruly.

4 Method 1.

Measurement of two and seven

days heat of hydration of cement by isothermal calorimetry (alternative to solution calorime-

try)
4.L fntroduction
The method described here is an alternative to solution calorimetry for the
determination of the heat of hydration of cement. It has many similarities
to the method proposed by Cembureau (1977) and it has the advantage over
solution calorimetry in that it can be used for all types of cements. It also gives
the thermal power during the whole reaction, and the same or almost the same
method may be used for many other purposes (cf. section 5 below). A difference
compared to solution calorimetry is that isothermal calorimetry cannot measure
the heat exactly from from the time of mixing. In the method proposed here
it takes about two minutes from the water is added to the sample is in the
measutement instrument. However, the heat produced during the first minutes
is very small compared to the heat produced during three or seven days (cf.
Appendix E).
The standard conditions for this method are as follows:

pure cement paste

o wf c:0.40
o hydration temperature: 20oC

4.2 Laboratory
The laboratory where the preparation of specimens take place shall be maintained at a temperature of (20+2)"C.

4.3

Device for mixing

Mixing should be made in a small transparent beaker of plastics or glass using

a mixing device of plastics or stainless steel.

4.4

Balance

The calibrated balance used should have a precision of 0.01 g or better

4.6

Sample vials

Vials of borosilicate glass sealed with Teflon coated rubber seals should be used.
Other types of vials (ampoules) may also be used as long as any reaction between
the cement paste and the sample or the thermal properties of the vial does not
influence the measured result.
3

4.6

Thermostated isothermal heat conduction calorimeter

The calorimeter should have a temperature stability over 24 h of better that

+0.3 K. The baseline stability should be better than 10 L,W over 24h.

4.7 Calibration of calorimeter

Each calorimeter shall be calibrated with an electrical heater placed either in
the sample position or on the sample holder. The calibration should be made
in such a way that the standard deviation of the calibration coefficient (e) from
a series of calibrations should be less than I% of the calibration coefficient.
The time constant (r) may be determined from calibrations with an old
cement paste sample with low reaction rate or calculated from masses and heat
capacities of the parts and materials used (Appendix H).
The baseline of each calorimeter shall be determined regularity.

4.8

Measurement procedure

The measurement is made in the following steps (the described procedure is for
a multi-channel calorimeter) :

1. Cement and water should have been stored in amounts of less than 1 kg
at (2012)"C for at least 24 h before the measurement.
2. The following items are prepared for the measurement: laboratory note
book, cement, water, mixing beakers, mixing devices, glass vials, caps
(and capping device if this is used), devices to charge samples into glass
vials (e.g. plastic funnels) and other devices necessary for the calorimetric
experiments (e.g. hooks for the vials).

3. Weigh up 1010.05 g cement in the eight mixing jars.

4. Start calorimeter data collection and count-up timer at the same time to
keep track of the time (minutes and seconds).
5. Tare an empty vial.
6. Add 4+0.025 g water directly to one cement (a calibrated automated
pipette is recomended).

7. Note time when water was added (minutes and

seconds).

8. Mix carefully for 30 seconds. Make sure that there is no dry cement on
the bottom of the ja.r.

9. Add cement paste to the tared vial.

10. Determine and note mass of cement paste. The mass of cement paste
should be 4-8 g.

11. Seal vial and load into calorimeter

12. Note time when vial was loaded into the calorimeter (minutes and seconds). There should be no more than 200 seconds between mixing and
charging.

13. Repeat steps 5-12 for all samples

A few notes on the above procedure:

o Note that one has to have separate beakers, mixing devices, funnels

etc.

for each sample.

It

is advantageous if the sample is as close to 20oC as possible when the

sample is charged. Therefore do not heat the sample with your hands.
Use cotton gloves.

o The calorimeter should not have a low upper range for this type of measurement as it then will take a long time before the thermal power goes
bIow the maximal level.2

4.9 Evaluationprocedure
For the evaluation one needs the calibration coefficient e (W/V), the time constant r (s) and the baseline Ur (V). It is here assumed that the primary output
from the calorimeter is in volts, but other r:nits may also be used (Appendix I).
The thermal power P (W) is calculated from the measured voltage U (V) with
the following equation:

p:e(U-u,+,#)

(1)

The dU /dt-part does only play a role at the very beginning of the measurement.
All diagrams should be made so that time zero is time of mifng. Atl integrations should be made from the time of charging of the sample into the
calorimeter (essentially the same as from the time of mixing as the thermal
power (cf. Eq. 1) should be zero (:baseline) before the sample is charged).
2The TAM Air has two ranges (60 mW and 600 mW); the upper ange should be used for

this measurement.

5 Method 2.

Measurement of heat production

rate of cement paste, mortar and micro concrete by isothermal calorimetry at different
temperatures (alternative to semi-adiabatic and
adiabatic calorimetry)

5.1

Introduction

The method described here is an alternative to semi-adiabatic and adiabatic

calorimetry for the determination of the heat of hydration of cement. It may
be used both with the standa.rd conditions given below and as a general tool to
investigate mortar/concrete hydration. It has many similarities to the method
proposed by Cembureat (1977). In the method proposed here it takes 2-5
minutes from the water is added to the sample is in the measurement instrument.

The standard conditions for this method as used as

this project are as follows:

in the

second

part of

cement mortar with an unreactive aggregate with a maximal diameter of

2 mrrr.

o w/c:0.50

o cement/aggregate-ratio

2.0

o hydration temperatures: 20, 30, 40, 50 and 60oC

Other condictions may also be used.

5.2 Laboratory
The laboratory where the preparation of specimens take place shall be maintained at a temperature of (20+2)"C.

5.3

Device for mixing

Mixing should be made in a small transparent beaker of plastics or glass using

a mixing device of plastics or stainless steel.

6.4

Balance

The calibrated balance used should have a precision of 0.01 g or better

5.5

Sample vials

Vials of borosilicate glass sealed with Teflon coated rubber seals should be used.
Other types of vials (ampoules) may also be used as long as any reaction between

the cement paste and the sample or the thermal properties of the vial does not
influence the measured result.

5.6

Thermostated isothermal heat conduction calorimeter

The calorimeter should have a temperature stability over 24 h of better that

+0.3 K. The baseline stability should be better than 10 L,W ovet 24h.

5.7 Calibration of calorimeter

Each calorimeter shal be calibrated with an electrical heater placed either in
the sample position or on the sample holder. The calibration should be made
in such a way that the standard deviation of the calibration coefficient from a
series of calibrations should be less thar. 7% of the calibration coefficient.
The baseline of each calorimeter shall be determined regularity.

5.8

Measurement procedure

The measurement is made in the following steps (the described procedure is for
a mu-lti-channel calorimeter)

1. Cement, aggregate and water should have been stored in amounts of less
than 1 kg at (2012)oC for at least 24 h before the measurement.
2. The following items are prepared for the measurement: laboratory note
book, cement, aggregate, water, mixing beakers, mixing devices, glass
vials, caps (and capping device if this is used), devices to charge samples into glass vials (plastic funnels) and other devices necessary for the
calorimetric experiments (e.g. hooks for the vials).

3. Weigh 20+0.1 g cement and 10*0.05

and mix them.

g flne aggregate in the

mixing beakers

4. Start calorimeter data collection and count-up timer at the same time to
keep track of the time (minutes and seconds).
5. Tare a vial

6. Add 5+0.025 g of water to one of the cementfaggragate mixes and mix

by hand for 30 s. Make sure that all cement is mixed. Note time when
water was added (minutes and seconds).

See note below on mixing.

7. Charge the tared vial with sample and determine the sample
8.

Seal

vial and load into calorimeter.

mass.

9. Note time when vial was loaded into the calorimeter (minutes and seconds). There should be no more than 300 seconds between mixing and
charging for the standard procedure.

10. Repeat steps 5-7 for all samples

A few notes on the above procedure:

o Note that one dhould have separate beakers, mixing devices, funnels for
charging etc. for each sample.

It is advantageous if

the sample is as close to the temperature of the

calorimeter as possible when the sample is charged (to decrease the time
that the thermal power cannot be determined). This is obviously more
difficult when measurements are made at non-ambient temperatures. One
may try to thermostat the different materials so that the mixed sample
has the temperature of the measurment when it is being charged into the
calorimeter (cf. Appendix J). Do not heat the sample with your hands.
Use cotton gloves.

o Mixing by hand is the standard procedure. There may be cements for

which longer mixing by a mechanical mixer is neccesary.

The calorimeter should not have a low upper range for this type of measurement as it then will take a long time before the thermal po\/er goes
below the maximal level.s

The maximal aggregate size that should be used in the TAM Air is about
5 mm diameter.

5.9

Evaluation procedure

For the evaluation one needs the calibration coeffi.cient e (W/V) and the baseline
Ur (V). It is here assumed that the primary output from the calorimeter is in
volts, but other units may also be used (Appendix I). The thermal power P (W)
is calculated from the measured voltage U (V) with the following equation:

e(U

- t/r)

(2)

It

is not neccesary to make a Tian correction.

AII diagrams should be made so that time zero is time of mixing. AII integrations should be made from the time of charging of the sample into the
calorimeter (essentially the same a from the time of mixing as the thermal
power (cf. Eq. 2) should be zero (:baseline) before the sample is charged).
sThe TAM Air has tv/o ranges (60 mW and 600 mW); the upper range should be used for
this measurement.

Various uses of isothermal heat conduction

calorimetry for the study of cement hydration

This section summarizes various applications of isothermal calorimetry for the

study of cement hydration. The range of applications is very wide and the list
below is by no means conclusive.

,A'.1 Determination of three and seven days heats of hydration

This measurement is made to compare cements and classify them according
to their rate of reaction (heat development rate). It is described in detail in
section 4. This type of measurement by isotherma calorimetry is an alternative
to solution calorimetry.

4.2 Heat of hydration

as a function of temperature

These type of measurements (cf. section 5) are made to assess the influence of
temperature on the heat production from cement paste, mortar and concrete.
Such information can be used to calculate temperature distributions in massive

constructions, and determine the time for deforming. Figures 1 and 2shows

measurements on a cement paste with w/c:0.5 at three temperatures. It is

seen that the thermal powers, and thus the rate of reaction, increases with
temperature.
In Fig. 3 the thermal power is plotted as a function of the produced heat.
This plot therefore shows the rate of reaction as a function of reacted amount.
It is seen that all three curves peak at the same time ard shows the same weak
second peak. The produced heat is a measure of the degree of hydration.

4.3 Determination

of admixture retardation

Most plasticizers retard the cement hydration and isothermal calorimetry is

an good method to assess the retardation time as a finction of temperature,
admixture concentration etc. Figure 4 gives an example.

^.4

The influence of contaminants on cement hydration

Water and aggregate can contain substances that influences the rate of cement
hydration. There have also been cases when cement trucks have been contaminated by other products, e.g. fertilizers, that influence the rate of hydration.
Figure 5 shows the result from measurements at 20"C on a mortar containing
different fractions of a mixture of soil and saw dust. The base recipe contained 50

(J

E5
OJ

20oc
30o

10(}c

E
L

oJ

E
lE

E
L
fJ
!

0
u

20

40

lime /

60

Figure L: Thermal power of cement hydration at three temperatures: 10, 20 and

30"C. The samples were cement pastes with wf c:0.5 For each temperature the
results of foru simultaneous measurements of samples from the same mix are
plotted.

10

350
300

250

fizon

E]

: 15
= 1oo
(E
(

50

50

100

time

10oc
20oc

3o
150

Figure 2: Heat produced from cement reaction at thee temperatures.

11

E5
)

c)

E
L

(r}

100c

200c

3o

.)
J

!-

EC.

{l
!

0
0

50

100

150

heat

200 25U 300

350

/ J/9""r"r,t

Figure 3: Thermal power (proportional to the rate of reaction) ploted as fwrction

of produced heat (proportional to the reacted amount).

72

o) 3

o
E 2
co
E
0
5

E
o_
d

0
0

10

15

20

25

Tme/h

Figure 4: This figure shows the result of measuements on six cement pastes
with different concentrations (0; 0.1; 0.15; 0.2; 0.3; 0.5%) of an admixture.

13

n contaminalon

- - low contamntion
medium contaminatinn
--'----

'

q
fJ

high contmination

'J

E
(lJ

c-

E
L

(lJ
!

2D

4t
time

E0

1t0

Figure 5: The result from a measurement on four cement mortars with different
proportions of a contaminant.

(w/c:0.6). To this mixture the contaminant

mixtnre was added in four concentrations: 0, 4, 12 and29% of the cement mass.
It is seen in the figure that the contaminant mixture retarded the hydration,
but that the hydration still continued at almost the same rate as without the
contaminant.Figure 6 shows the produced heat as a function of time. Here it is
evident that there is a slight delay in the reaction. Note that the calorimetric
g cement, 150 g sand and 30 g water

measurement shows the reaction rate which for an unknown concrete (especially
with contaminants like sawdust present) does not give any information about
t}ne absolute strength development that is also a function ofstructure, aggregate
etc.
This example shows that isothermal calorimetry is an interesting method to
judge the influence of contaminants on cement hydration, whether the purpose
is to judge if a certain aggregate is acceptable or find the source of a problem.

4.5

Studies of early reactions

Directly after water has been added to cement reactions stat. All applications
described above refer to the main reaction which at 20"C starts at about 3 h
after mixing. To study the earlier reactions there are two options:

I4

250

200
E
itl

{J

G)

tso

fD
!
1

(I)

00

---'-'--'

50

0
0

2t

no contamination
low contamination
medium contamination
high contamination

40

6t

time

Figure

15

6:

80

100

Charge directly after mixing and correct the signal from the calorimeter
with the Tian equation (maybe with two time constants).

Charge only the cement into a vial connected to a mixing device. Inject
water when cement is at equilibrium and mix the sample inside the vial
while it is in the calorimeter. Correct the signal with the Tian equation
(maybe with two time constants).

A comparison between different types of calorimetry for cement hydration studies

8.1 Isothermal calorimetry

The heat production in the sample is monitored while the sample temperature
is unchanged. The most common type is the heat conduction (flow) calorimeter
is which the heat produced in the sample is measured as it is conducted out

from the sample through heat flow meters. There are today several types of
commercial instruments available. No cement standards exist for isothermal
calorimetry, although Cembureau have issued recommendations how to use it
(Cembureau 1,977) and there have been a few manufacturers of such instruments.

8.2 Adiabatic calorimetry

In this type of calorimetry the

sample is perfectly insulated so that all heat

produced is used to heat the sample. The temperature increase of the sample is
measured and the rate of temperature increase is a measure of the heat production rate in the sample. So called adiabatic shields are usually used to hinder
the heat form leaving the sample.

El.3

Semi-adiabatic calorimetry

This is a simpler type of insulated calorimeter in which the insulation is not

perfect so a certain fraction of the heat produced is allowed to escape from
the sample. In practice the sample is simply placed in an insulation made of
polystyrene. One example of a semi-adiabatic calorimeter is "hkassen" (NT
Build 388).

8.4

Solution calorimetry

In this method the total heat of hydration at a certain time is determined

as

the difference between the liberated heat when an un-hydrated sample and the
sample unde investigation is dissolved in a mixture of hydrofluoric acid and
nitric acid. This old measurement technique is described in ASTM C185, prEN
196-8, SS B1 1960 etc. It is time consuming, costly and dangerous, but still in
use.

16

8.5

Comparison between isothermal and adiabatic calorime-

try

Isothermal calorimeters have many advantages over adiabatic calorimeters:

o Isothermal calorimeters directly measure the heat production rate that is

proportional to the rate of the reaction (adiabatic calorimeters measure
temperature (change) that is recalculated to give heat produced).
o In isothermal calorimetry it is not necessary to know the heat capacity of
the sample, as is required in adiabatic calorimetry.
o Isothermal calorimeters are very stable and need not be calibrated more
than a few times a year (adiabatic calorimeters are often calibrated before
each run).

o Isothermal heat conduction calorimeters for

cement hydration studies are

more sensitive than adiabatic calorimeters.

o The temperature

never increases to unrealistic temperatures in an isothermal calorimeter. (final temperatures during measurements in adiabatic
caloimeters are very high, e.g. 80'C; far above what is desirable in real

constructions).

o The

engineer does not need adiabatic heat curves, but instead heat curves
for different constant temperatures as this is the information used in calculation of maturity and strength growth in concrete, as well as the risk
of thermal cracking.

Most calorimeters in use in cement and concrete science are instrurirents that
are either home built or manufactured only for this special use. The type of
modern isothermal calorimeter we plan to use in this project is commercially
available and is used in pharmaceutical industr microbiological laboratories,
defense research establishments etc. It is therefore a less expensive instrument.
Isothermal calorimetric experiments are easy to perform and need little operator
time (in contrast especially to the conduction calorimetic method described on
the previous page).
The only possible dawback with isothermal calorimetry is that only small
samples may be used. For measurements on large samples of concrete one
therefore needs adiabatic or semi-adiabatic calorimeters. Most measurements
may however be made on cement pastes, mortar or micro-concrete since normal
coarse aggregate does not influence the heat development in the cement paste

(approx. 30% of the concrete). Calorimetric measurements on a real concrete

could also be done isothermally by sieving the large aggregate away before the
measurement.

Adiabatic and semi-adiabatic calorimeters are sometimes advocated because

the temperature increases during a measurement, just like it does in a real construction. This is, however, not of any advantage since there are very different

77

temperature increases in different parts of a massive construction. The truth is

that the isothermal measurement is the most well defined calorimetric measurement, and to completely describe the hydration of a cement paste or a mortar
the best way is to make isothermal measurements of its hydration at different
temperatures.

C The influence of aggregate on cement hydration

As noted in Appendix B it is not possible to have large aggregate in the small
samples for isothermal calorimetry. The maximal aggregate size for a TAM Air
20 ml glass ampoule in the order of 5 mm, but one should note that the larger
the aggregate is, the larger will also the spread in the result be. In cement
calorimetry we want to measure the rate of cement hydration. Do we need to
have the aggregate in our calorimetric samples? Would it be possible to make

calorimetric measurements only on cement paste or only on mortar (with a

maximal aggregate size of 2 mm)?
How can the large aggregate influence the hydration? I see three possible

'ways:

If the aggregate is dry it may absorb water

lowering the water-cement

ratio. It may also be wet so that it raises the water-cement ratio. Normal
Swedish aggregate has a quite low porosity and small changes in watercement ratio will not influence the cement hydration much.

o It carries on its surface contaminants that influences the hydration. If one

suspects that this is the case the water that shou-ld be used for mixing the
mortar could be used to leach out these contaminants prior to the mixing.

o It is extremely reactive, but such aggregate should not be used.

A number of tests on the influence of the aggregate on cement hydration

has been performed as part of this work. The idea was to see if there was a
difference between no aggregate and fine aggregate as it can be expcted that
the fine aggregare with its larger surface area should influence the hydration
more than the large aggregate. Different types of small size aggregates (sand,
gravel) were used. All measurements were made with Cementa Byggcement
with a water-cement ratio of 0.5. For each measurement a cement paste was
mixed and divided into three parts. One was charged into two ampoules as
it was (cement paste), to one part an equal mass of aggregate was added to
the cement paste, to the last part half the mass of aggregate was added to the
cement paste. The two mortars were charged into three ampoules each. It
should be noted that the second type of sample with equal masses of aggregate
and cement paste was very diff.cult to mix into a mortar by hand; it contained
an extreme fraction of fine aggregate.

18

3.5

--

'-'--'

OJ

'

cement paste
max <t.25
half max <0.25

2.5

')

t,s

U
a-

E1
t
L
(J

o.s
0
0

50

100

time

15t

Figure 7: All results from one measurement with the same cement paste, but
very different amotnts of aggregate (fine sand).

19

u.9

t
t

o,
E
0

t
\ 'rt

*-

-'

:'

E
L

ol

0 7

E 06
(I

-"-"--

0.5

0.5

1.5 2
time

2.5

cement paste
max <t.25
half max <0.25

3.5

Figure 8: Close up of start of the main cement hydration. The same data
shown in the previous figure.

20

as

tr

{,}

,!

E}

2.

tt';!

lt
(J

2. 6
o
o-

It

Ec
L

{t

24

2.2
4

14

81012
time /

16

Figure 9: Close up of the peak of the measurement shown in Fig.

2t

Figures 7-9 shows typical results from these measurements. Figure 7 shows

that the results are rather similar for both cement paste and mortar. Figure 8
shows that the main reaction in the cement paste starts about 20 minutes earlier
in the cement paste that in the mortars. However, in another measurement the
cement paste hydration started about 12 minutes later than the hydration of
the cement mortars, so this small difference seems to be variable. Figure 8
also shows that the thermal activity during the dormant period is higher in the
mortars that in the cement paste, but as this is during a relatively short time it
does not much influence the total heat. Finally, Fig. 9 shows that the maximal
thermal power is slightly higher for the mortars than for the cement paste. It
also shows that there is a second peak for the samples with the largest fraction
of aggregate.
The differences seen between the samples with and without aggregate are
quite small. This suggests that one may determine the heat production rate of
a concrete by making measurements on cement paste or cement mortar with
only file aggregate. Table 2 in Appendix F also shows that the heat integrals
are about the same with and without aggregate.
For the large majority of cases I believe that it is safe to make measurements
on cement mortar containing aggregate with a diameter of 2 mm or less. The
heat production per unit ofmass ofconcrete also containing the larger aggregate
is then:
n P rn. I rn* +Tnurlf

n'Ls

'trL.

T|t*

ntr^o

(3)

Here, P" (W) is the measure thermal power from the sample, -" (s) is the mass
of the calorimetric sample, m" (S) is the mass of cement used in the mix, rn*
(g) is the mass of water use in the mix, rn,, (g) is the mass of the agregate
used in the mix, / is the mass fraction of the aggregate that was used for
the calorimetric sample, and II (W/g) is the specific thermal power from the
mortar/concrete sample with all the aggregate. \Mhen a sieve is used to sort
out the larger aggregate from the smaller that can be used in the calorimetric
sample, then / can be calculated like:

ffL^"""-,t
?TLpassed

Here, Dasse (S) ir the mass that passed the sieve and moo p".."a (g)
mass that did not pass the sieve.

(4)

?7not passed

is the

Can glass vials be used for cement hydration

measurements?

Cement paste is extremely alkaline and can dissolve (etch) glass. I have therefore
made measurements with cement paste and cement mortar in glass vials and in
polyethylene bags to see if there is a difference in the measured thermal power.
Three types of samples were run:

22

(J

--

t.5

in vial

inPEbag

OJ

E
OJ

sE

o_

E 05
E

0
0

2t

40
time

60

8U

100

Figure 10: The result of measurements of cement paste hydration. Four samples
directly in glass vials and four samples in polyethylene bags with parafin oil as
heat transfer medium.

1. The samples were charged directly into the vials.

2. The samples were charged into polyethylene bags that were placed in glass
vials. This did give somewhat lower thermal powers and a third type of
samples were then used to see if this could be caused by the limited therrnal
contact between the plastic bag and the glass vial.

3. The samples were charged into polyethylene bags that were placed in glass
vials. About 1 ml of parafrn oil was placed as a heat trarsfer medium on
the bottom of the vial. This type of sample gave the same result as did
the samples charged directly into vials.

In the following only types 1 and 3 are discussed. Figure 10 shows the resu-lt
of measurements with a cement paste of water-cement ratio of 0.5. Table 1 gives
evaluated heats that shows that there is no difference between measurements in
vials and in polyethylene bags. It is therefore possible to charge cement paste
and cement mortar directly into glass vials for measurements.

OD

IJ
748.0+2.7

heat

in glass vial
in polyethylene bag

hear

746.7+1..2

746.7+r.0

Table 1: Mean values and standard deviations of the integrals of thermal power

from 140 minutes after mixing (start of main reaction) to SO00 minutes. In
the rightmost column one of the samples in a glass vial has been excluded as
it shows a slightly different shape of the hydration curve (the one with more
pronounced second peak in Fig. 10).

Integration limits for heat of hydration

A most common calorimetric measurement for cement is the determination of

the heat of hydration during the first three or seven days of hydration. For the
three types of calorimetric methods used (cf. Appendix B) different integration
Iimits must be used:

o For isothermal calorimetry

as described in this report a cement paste or a

mortar is mixed outside the calorimeter and then transferred into the
calorimeter. Depending on the circumstances the time from the start
of the reaction (addition of the water to the cement) to the charging
of the vials into the calorimeter can be from two to 30 minutes for
an isothermal calorimetera. The longer time refers to a case where
several different pastes are mixed separately, but charged into the
calorimeter at the same time. If one is interested in knowing the
true heat from the early reactions one should of course work with
short times between mixing and charging. After a sample has been
charged into the calorimeter it will take in the order of 30 minutes
before the true thermal power is directly measured. This time can
be substantially shortened by the use of the Tian equation which
calculates the true thermal power in the sample from the measured

signal.

o For semi-adiabatic and adiabatic calorimeters the sampe is usually

large and mixed outside the instrument and then transferred to the
instrument. The time for this is very much dependent on the type of
instrument. It will then take some more time before the sample has
thermally stabilized itself. All in all the time between mixing and
knowing true values is usually in the order of 30-60 minutes.

o The solution calorimetric method is quite different from the above

methods as one only gets information about the produced heat from
time zero (start of mixing) to the time when one ends the measurement. This is thermodynamically very elegant, but the measurement
the TAM Air procedures in sections 4 and 5 it should take approx. two minutes to

aFor
get the sample into the calorimeter.

24

07

hn
5
-'-5)
tI}

--

in vial

inPEbag

U.Ti

{D

0.55

(g

tc

o)

fr5

E,

0 45

2.4 2.6 2B 3 32 3.4

time

3.6

Figure 11: A close up of the start of the cement hydation in Fig.

11

is difficult to make and one does not get any information on what
happens during the time until the measurement is stopped.
For isothermal calorimetry there is a possibility to integrate from the start of
the main hydration as this time is clearly seen as a knick-point on the diagram,
as in Fig. 11. Here, it is clearly seen that all eight measurements have this knickpoint at about 3 h. I have defrned the starting time of the main hydration as the
inflexion point (cf. Fig. 12).Figure 13 shows a results from a measurement in

which the sample was charged about two minutes after mixing. The dashed line
(with Tian correction) is show a somewhat higher total heat than the solid line
(without Tian correction). This is because the Tian correction takes away the
heat in the ampoule that is there because it did not have the same temperature
as the calorimeter when it was charged. The logarithmic scale makes this most
visible at the start of the measurement, although the difference exists for the
whole measurements (after the initial part the thermal power changes so slowly
that the Tian equation does not make much difference).

Reproducible mixing

The mixing procedures prescribed in cement, morta and concrete standards

are rather detailed. However, I have found that for the small mixtures (10-100
25

0.55

c)
t)

E
L

(l}

u5

3
o

E
E
L
ft)

0.45

2.6

2.8

3
time

3.2

34

3.6

Figure 12: The arrows shows how I have defined the starting time of the main
cement reaction as the inf.exion point.

26

10

10

wlthout Tian
wlth Tlan

ol

o
o

mtn

0h

day

days

J
E

o
o0

10

-1

-2

10

l0

-1

100

101

limo /

10

103

Figure 13: The result from a me,rurement with quick ctrarging in a logJog
diagram to show the differences at early times (not that the scale grossly overexagerates the differences in the beginning of the measurement).

27

mortar wf c:0.5
paste

w/c:0.5

three mortars w/c:0.5

one paste wf

c:0.5

strt react

Q, start-10 h

2.92+0.07
3.03+0.05
2.55+0.01
2.33+0.02

33.2+0.4

/9."^.n

32.1+1.0
50.0+0.2

49.3r0.2

,start-3days

main peak

J/""*"rt

mW/g""m"nt

198+4
206+6

1.64+0.03
1.65+0.05
2.50+0.10
2.50+0.004

257.6+6.6
269.9+2.4

Table 2: Results from two mixes with the same vvter-cement ratio. All figures
are mean and standard deviation of four measurements.The three mortars on
Iine three were mixed by hand 15 s, 1 min and 2 min.

g) of cement paste or mortar that is used in the methods described above one
minutes mixing by hand gives reproducible results. The last two lines of Table
2 show one example of this (the top two lines also show this as the mixing of the
cement paste is supposed to be much more efficient when there is an aggregate
present as in the mortar).
There may be cements with high flller concentrations etc. for which hand
mixing is not suffi.cient. One may then use small electric mixers (kitchen type).

Basics of isothermal calorimetry for cernent

hydration
When cement is mixed with water a series of chemical reactions starts, ending
with the hardened cement paste. All reactions, including the ones involved in
cement paste hydration, produce heat. The measurement of the heat produced
by cement hydration by isothermal calorimetry is therefore a method to monitor
and study the hydration processes.
An isothermal calorimeter measures the thermal power (heat production
rate) P (unit: \M) from the cement hydration reaction. Let us assume that the
aggregate does not influence the hydration reaction. Then for a cement paste
sample with a water-cement ratio w f c (g/g) and a mass nz (g) the thermal
power is proportional to the enthay of hydration Ln Q/ and the degree of
hydration o:

dl
P:
- Lh+.
wlc+ 1 d

(5)

Under isothermal conditions and for a certain cement paste the thermal power
is a function of time or the degree of hydration (either may be used):

p: f{t): fz,)

(6)

Under non-isothermal conditions the thermal power can only be described as a

function of temperature ? (K) and t}re degree of hydration. This is done by
assuming that the thermal power at a certain degree of hydration follows an

28

Arrhenius'relationship:

_p

P: Po(a)exp(;ff)

(7)

Here, ,. could be a function of a, but one usually takes this as constant.

The degree of hydration can be determined in different ways. For the present
type of measurement it is useful to define it as the heat of hydration (J) dividea
by the maximal heat of hydration (when all cement has reacted):

":ffia

(8)

The heat is the integral of the thermal power:

Q(r):

Pdt

(e)

An isothermal calorimeter measures thermal power which is proportional to

reaction rate. From such a mea"surement we can therefore see how the cement
hydrates. This is useful for many different purposes, as discussed below. However,

it

should be noted that there are t'wo types properties of which isothermal

calorimetry 'in general does not give information about: rheology and strength.
It should be noted that Ah in the above equations is not a constant. It
will change during the course of the cement hydration and it will be he weighed
mean of the Ah,:s of all different types of reactions taking place simultaneously.

IJse of the Tian equation for cement hydration studies

The measured signal from a heat conduction calorimeter is not the true thermal
power produced in the sample as there is a time lag for the heat to pass through
the heat flow sensor. The Tian equation removes this time lag and gives as
output the true thermal power produced at each instant in the sample:

: e(U * r#l

(10)

Here, P (W) is the thermal power, U (X) is the measured output, e (W/X) is
the calibration coefficient, a.nd r (s) is the time constant. The measured output
can have different units - here denoted by X - volts being the most common,
but the output from an A/D-converter or mm measured on a recorder paper
could also be used.
It is unnecessary to use the Tian equation if the process studied only changes
slowly compared to the time constart r which may be calcu-lated by the following
equation:

'-k
29

11)

C,JIK
ampoule holder

72

glass ampoule
cement

13

c, J/Klg
0.76
4.18
0.75

water
aggregate

Table 3: Thermal properties for time constant calculations.

Here, C (UK) is the total heat capacity of the sample (including ampoue and
ampoule holder etc.) and k (W lK) is the thermal conductance of the heat flow
sensor. The heat capacity C may be calculated by summing the heat capacities
of all r parts on the sample-side of the heat flow sensor:

c:icn
i:t

02)

For cement paste and mortar in the TAM Air with glass ampoules the values
in Table 3 may be used. The thermal conductance of the TAM Air heat flow
sensors is 0.68 W/K. The time constant may also be found from dynamic
calibration curves.

Figure 14 shows an example of a Tian-corrected curve measured on 10 g

It shows that part of a cement paste hydration curve where the
most rapid changes take place except for the very early reactions. The difference
is small and it is safe not to use the Tian correction for most applications. The
sample in this example has a time constant of about 65 s. One drawback with
using the Tian equation is that one has to collect data at a quite high rate
(minimum 5 s time interval). If one does not use the Tian correction any
initial disturbances can be considered to be gone after about 5r (7% of initial
distubance left), 7r (0.1%),9r (0.07%). 5
For the above example (Fig. 14) the integrals from the time the samples are
charged into the calorimeter (about 25 minutes after mixing in this case) to are
315.15 and 314.93 Jf E"" .ot, respectively, without and with the application of
the Tian equation. Starting the integration at a later time gives somewhat lower
heat integrals, e.g. 309.72 and 307.52 Jf gfor integration from 1 and 2 hours
a"fter charging. The conclusion from this is that if the goal of a measurement is
to look at the main hydration it is not important to use the Tian equations. On
the other hand, if one is interested in the integral (cf. the solution calorimetric
method) it is important to start measuring as early as possible from the mixing,
use the Tian equation, and start the integration as early as possible.
cement paste.

SFor cement hydration

50

with 5 g of cement paste in a TAM Air the time constant is about

s.

30

0.65

c
o)

o
o

o)

0.6

o)

3
o

o_

b
o

-c

0.55

1.6 1.7

1.8

1.9

time

2.1

2.2

Figure 14: An example of Tian correction. The dashed curve is the original
measured curve and the solid curve is Tian corrected.

31

Calibration of isothermal heat conduction calorimeters

Today the only simple and precise way of calibrating isothermal heat conduction
instruments for cement hydration measurements are by electrical calibration. In
such a calibration an electrical current is passed through a resistance (heater)
placed in the calorimetric ampoue or on he ampoule holder. If the current 1
(A) and the resistance is .R (f-l) the produced thermal power (P, W) can be
calculated as:

P: I2R

(13)

If the corresponding output from the calorimeter at steady-state is L/" (X) the
calibration coeffi.cient e (W/X) is calculated by:
P
e:fr

(14)

Here, X denotes the units of the data logger (usually V, mV, L,Y etc, but it can
also be the direct output form an A/D-converter).
Electrical calibration is good, but the results can be several percent wrong

if proper precautions

are not taken. The main rule is to make the calibration

case as similar as possible to the measurement case. Place the heater as close as
possible to the place where heat will be produced during a measurement. Use
thin electrical leads (but not too thin as this will give heating in the leads).

J A method to start a measurement at the correct ternperature

It

is a problem, especiall if one is working with calorimeter at another temperature than the laboratory, that the sample does not have the same temperature
as the calorimeter when it is charged. Below is presented a method to calculate the temperature of the water so that the flnal mix is quite close to the
temperature of measurem"t. ttr" used nomenclature is according to Table 4.

The mean weighted temperature of all components (incl. vial etc.) should
be equal to the temperature of the calorimeter:

T*m*c* jTomuc^+TuC*
mwcwlm^cnlC*

(15)

:fc

This may be rearraged to:

T*:

T"(m*c*

nluco

+ C*)

Tumucu

mrycw

If water with this temperature is used to mix the

+ ToC*

(16)

cement paste of mortar the

temperature of the mix should be similar to that of the calorimeter.
to

cs
qv
C*
rrtrs

mN
Tu
f.
T*

specific heat capacity of cement and aggregate 0.75 JlglK

speciflc heat capacity of water
a.2 J lglK
heat capacity of vial and ampoule holder6
J lI<
mass of solids (cement, aggregate)
g
mass of water

in

pase/mortar

temperature in lab (temp. of solids)

temperature in calorimeter
temperature of water in cement/mortar

g
oC
oC
oC

Table 4: Nomenclature used in method to calculate temperature of water when

mixing paste and mortar.

K flow isothermal is an isothermal heat conduction calorimeter?

Isothermal heat conduction calorimetry is essentially isothermal, not exactly
isothermal, as the heat produced in the sample will heat the sample slightly
so that one gets the temperature gradient that gives the conduction of heat
away from the sample. Wads ("Temperature changes within samples in heat
conduction calorimeters" Thermochim. Acta 366 (2000) 72I-127) has given
equations to calculate the maximal (center) temperature increase in a sample
during a measurement. For most purposes the following equations may be used:
p

Lr:

(17)

Here, A? (K) is the temperature difference between the heat sink and the sample, P (\M) is the produced thermal porver, and k (W/K) is the thermal conductance of the heat flow sensorsT. For well designed measurements in well designed
isothermal heat conduction calorimeters the temperature increase should be less
than 0.1 K, i.e. the measurement ca.n be seen as essentially isothermal.

Research needs

Below are listed a number of questions that should be looked into to gain a
better understanding of isothermal calorimetry as a tool for the study of cement
hydration.

L.l Hydration as a function

of wf c

Water-cement ratio (w/c) is the most important parameter in concrete production as it greatly infl.uences both strength and rheology. However, its in-fluence
on the hydration rate is not at all as marked and one wouLd probably find that
TFor the TAM Air k:0.33 W/K which gives a temperature increase of approx. 0.05 K with
a 5 g cemenl paste sample.

d
'D

using slightly different w/c (e.g. 0.55 instead of 0.50) one would get essentially
the same results from a calorimetric measutement.
There are a number of factors of interest here:

o For normal w/c there is more water than the cement needs to fully harden.
Low wfc pastes are self desicating as there is not enough water proesent
for a full hydration of the cement.
o More water means that the cement particles are further away from

each

other.

o The surface of a concrete is often exposed to more water from watering

and rain than the interior.

It would be interesting to make measurements of hydration as a function of

water supply (incl. different w/c and wetting of surfaces). A combination of
isothermal calorimetry and RH-measurements could be of interest.

L.2 Internal mixing vs. external mixing for the study of

early reactions
It

is possible to mix cement pastes and mortars down in the calorimeter, at least
at high w/c. This should be investigated further.

L.3 Relative strength

development as a function of pro-

duced heat

It has been shown that there is a rather

linear relation between the strength

development and the produced heat. However, it is not known when this relation
starts as there can be no such relation before the cement paste has any strength,
i.e. before the main reaction starts, even if some heat is produced.

Practical questions to be answered in the second part of this project

1. How important is it to measure the heat production rate from the time of
mixing when one is interested in the temperature of a massive construction? Problems with measurements at non-ambient temperatures.

2. How important is the comntrol of the water-cement ratio?

3. How important is the contro of the mixing

process?

4. How important is the possible influence of aggregate on cement hydration

(moistrue content, porosity)

5. What precision, reproducibility is needed for isothermal calorimetry?

34